Low-Cost Diagnostics: Using Paper as a Material and Pens ...

Brigham Young University Brigham Young University

BYU ScholarsArchive BYU ScholarsArchive

Undergraduate Honors Theses

2020-06-16

Low-Cost Diagnostics: Using Paper as a Material and Pens as an Low-Cost Diagnostics: Using Paper as a Material and Pens as an

Instrument Instrument

Annie Armitstead

Follow this and additional works at: https://scholarsarchive.byu.edu/studentpub_uht

BYU ScholarsArchive Citation BYU ScholarsArchive Citation Armitstead, Annie, "Low-Cost Diagnostics: Using Paper as a Material and Pens as an Instrument" (2020). Undergraduate Honors Theses. 145. https://scholarsarchive.byu.edu/studentpub_uht/145

This Honors Thesis is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in Undergraduate Honors Theses by an authorized administrator of BYU ScholarsArchive. For more information, please contact [email protected], [email protected].

Honors Thesis

Low-Cost Diagnostics: Using Paper as a Material and Pens as an Instrument

Annie Armitstead

Submitted to Brigham Young University in partial fulfillment

of graduation requirements for University Honors

Department of Physiology and Developmental Biology

Brigham Young University

June 2020

ACKNOWLEDGMENTS: I would like to thank my husband, Jamon, for asking me

questions and helping me discover answers during this project’s lifetime. It was

ultimately he who thought of the simplest piece of this work: the antibody pen. I am

grateful to my father, Kelly, for lending me his social capital and the genesis of this

project, as well as many meetings to offer guidance and counsel. Without him, this

project would have not begun. Thank you to my mother, Emily, and my brother, Ethan,

who have been both good listeners and encouraging supporters when I’ve talked about

and presented on this research.

I would like to thank my advisor, Richard, for not only helping me to bring this

project to its full potential, but for helping me to become an independent scientist. Thanks

to him, I learned how to manage a world of unknowns while having fun exploring and

creating solutions. I am grateful for the positive magnet he is, and especially for the

people he drew towards him, whom I’ll also thank. Thank you to Lara, who became my

true friend through thick and thin, who helped me turn impossibilities in to possibilities,

and was the perfect counterpart in this research. Thank you to Kealani, a fast friend

whose drive and energy put new life into our project, taking it to a higher level of

achievement. Thank you to Taylor and Preston, for seizing this research in its second

wave, and being the kind of researchers to make it a legacy.

ABSTRACT: Lateral flow immunoassays (LFIs) are equipment-free tests that produce

results quickly using small sample volumes. Colored lines appear as the test runs,

indicating the presence of a biomarker. LFIs are ideal in a variety of settings.

Development of these assays can be complicated for small team operations, and tests are

not sufficiently adaptable for low resource settings. If robust point-of-care tests can be

developed on site, they can expand the reach of global diagnostics, improving health

around the world. We report a simple empowering LFI (seLFI) that only requires treated

printer paper and a plastic backing card. This eliminates the use of nitrocellulose

membrane, sample pad, conjugate pad, absorbance pad, and all the required layering

during assembly. Additionally, test and control line antibodies can be applied to this

treated paper by a rollerball pen, allowing on site preparation of the seLFI. The seLFI

is a low-cost, labor-saving, single-sheet diagnostic that performs comparably to

commercially available LFIs. These tests could assist small team operations in

developing versatile tests to use on site, or to resolve complications during

commercialization. They could also be used in low-resource or at-home settings to

personalize tests before running them. We present the proof-of-concept seLFI to improve

the accessibility, robustness and versatility of diagnostics research.

TABLE OF CONTENTS Introduction………………………………………………………………………..……..11

Methods……………………………………………………………………………..……13

Progression to Sandwich Model…………………………………………….…...13

Nitrated Printer Paper……………………………………………………….…...14

Aldehyde-functionalized Printer Paper……………………………………….….15

Antibody Pens……………………………………………………………………16

Testing in Urine………………………………………….…………………….…17

Testing in Serum…………………………………………………………………17

Limit of Detection…………………………………………..……………………17

Results……………………………………………………………………..……………..18

Nitrated Printer Paper…………………………………………………..………..18

Aldehyde-functionalized Printer Paper……………………………….………….19

Antibody Pens……………………………………………………………………20

Testing in Urine…………………………………………………………………..21

Testing in Serum…………………………………………………………………22

Limit of Detection……………………………………………………..…………23

Discussion……………………………………………………………………….……….24

References………………………………………………………………………………..26

LIST OF FIGURES

Figure 1: Traditional LFI to seLFI………………………………………………………...2

Figure 2: Antibody Pen…………………………………………………………………..16

Figure 3: seLFI with Nitration Method…………………………………………………..19

Figure 4: Aldehyde-functionalized seLFI Strips…………………………………………20

Figure 5: Antibody Pen Precision and Efficacy………………………………………….20

Figure 6: Antibody Pen Extinction Test………………………………………………….21

Figure 7: Aldehyde-functionalized seLFI Strips in Urine……………………………..…21

Figure 8: Aldehyde-functionalized seLFI Strips in Fetal Bovine Serum………………...22

Figure 9: Aldehyde-functionalized seLFI Limit of Detection…………………………...23

11

■ INTRODUCTION

Across the globe, regardless of

age, race, and condition of living, people

struggle to overcome or live with health

challenges. According to the World Health

Organization, the number of patients with

diabetes has risen to 422 million, with 3.7

million deaths caused by high glucose per

year. The World Health Report indicates

that 50,000 people die of infectious

diseases daily, many of which could be

prevented or cured for just one dollar a

person. Approximately 76,000 women

and 500,000 babies die worldwide from

preeclampsia and hypertensive disorders

yearly. In developing countries, women

are seven times more likely to develop

preeclampsia than women in developed

countries. The World Allergy

Organization estimated in 2011 that 30-

40% of the world’s population

experiences an allergy to one or more

allergens. In a world of chronic disease,

infectious agents, allergies, intolerances,

and health complications, the need for

diagnostic tests is becoming ever greater.

The World Health Organization has

developed a standard for diagnostic tests

called the ASSURED criteria: Affordable,

Sensitive, Specific, User-friendly, Rapid

and robust, Equipment-free, and

Deliverable to end-users. Diagnostic tests

following these criteria allow healthcare

providers with any range of resources to

participate in improving the health of their

patients.

In many situations, lateral flow

immunoassays, or LFIs, are used to

satisfy the ASSURED criteria, bringing

diagnostic tests to even the most resource-

restricted clinics. These equipment-free

tests produce results in 10-20 minutes and

use small sample volumes. LFIs are easy

to run, and the results are straightforward

to interpret; colored lines appear as the

tests run, indicating the presence of a

biomarker in a sample. Although

traditional LFIs seem ideal for low-

resource settings, additional challenges

remain. By definition, low-resource areas

struggle with poverty, and maintaining a

strong medical front against epidemics is

inherently expensive. While LFIs are

manageable for first-world patients and

clinics to purchase, affording these pre-

assembled products can be difficult for

small team operations globally.

In any setting, a diagnostic test

needs to be non-invasive and fast,

delivering quick results at the point of

care (POC). In low-resource settings, a

clinic might be without advanced training,

equipment, or electricity, so it is critical

that tests are easy to understand, run, and

interpret. To reduce costs in low-resource

settings, the tests also need to be easy,

straightforward, and inexpensive to

assemble; lowering the manufacturing

costs reduces the cost of the test. To

manufacture traditional LFI tests, a well-

equipped lab would have a striper (a

machine that immobilizes antibodies to

nitrocellulose membrane), a laminator (a

machine that aids the assembly of the

layered lateral flow cards), and a test

shear (a machine that cuts laminated cards

into individual tests). A striper costs

anywhere from $6,500 to $46,500

(ClaremontBio and BioDot), along with

$6,100 to $8,625 for a laminator (BioDot

and Kinematic Automation), and $13,600

to $16,790 for a test shear (BioDot and

Kinematic Automation). On average, the

start-up cost of a small-scale lateral flow

research lab is $47,520 for this equipment

and materials like nitrocellulose

membrane, plastic backing cards,

conjugate, sample and absorbance pads.

For small team operations,

especially those new to diagnostics

research, this up-front cost that is almost

12

too great. For labs and clinics in low-

resource settings, this equipment is too

expensive, leaving them to source

diagnostic tests from larger companies

that were able to afford the start-up cost.

Making assay development and diagnostic

test manufacturing more accessible to

these groups would promote good health

practices, research, and scientific standing

in these areas.

We introduce a new LFI platform

that lowers the barrier of assay

development and manufacturing for low-

resource labs and clinics, as well as for

small team operations. This platform,

referred to as the “simple empowering

lateral flow immunoassay” (seLFI), uses

chemically treated paper attached on a

plastic backing support, reducing the

many layers of a traditional LFI to a

single sheet of paper. Like the traditional

LFI, the seLFI uses capillary action,

conjugated antibodies, test and control

lines, and a greatly simplified chase buffer

to detect a biomarker.

We report the development of the

seLFI, an LFI whose manufacture only

requires treated printer paper and a

supportive backing. This eliminates the

need for nitrocellulose paper, sample pad,

conjugate pad, absorbance pad and all

assembly associated with these layers.

Additionally, test and control line

antibodies can be applied to this paper by

an antibody pen, making the seLFI both

adaptable and straightforward to develop.

The seLFI is a low-cost, labor-saving,

single sheet of treated paper that produces

results comparable to commercially

available LFIs. In this paper, we describe

a seLFI test developed to detect hCG in

different sample matrices (spiked human

urine, fetal bovine serum) at levels of

1:103 in under five minutes. These tests

could assist individual researchers and

small team operations in developing

personalized, on-site diagnostics. We

present the proof-of-concept seLFI to

improve the production, robustness, and

versatility of diagnostics in small team

and low-resource operations.

A traditional LFI is made up of

several layered materials adhered to a

plastic backing card (Figure 1). These

materials consist of a sample pad, a

conjugate pad, a nitrocellulose membrane,

an absorbance pad, and often require tape

allowing compression between layers to

facilitate flow (Fenton, et al.). Each of

these layered materials are customizable,

Absorbance pad

Nitrocellulose membrane

Plastic backing card

Control line

Test line

Conjugate pad

Sample

Sample pad

Control lineTest line

Treated printer paper

Plastic backing card

Traditional LFI

seLFI

Figure 1. Traditional LFI to seLFI. The multiple layers of an LFI combine to allow

the flow of a sample through the diagnostic. The sample pad wicks up the sample, delivering it to

the conjugate pad. Here, the sample interacts

with the conjugated antibodies in the pad, and

any biomarkers present bind to the conjugated

antibody. The whole sample continues through

the nitrocellulose membrane, where any

biomarker/conjugated antibody complexes bind

the test line and left over conjugated antibodies

bind to the control line. The absorbance pad

ensures the complete flow of the sample through

the test, drawing it continually upward. The

seLFI removes all of these layers, relying on the capillary action of treated printer paper to draw

the sample, which is previously mixed with

conjugated antibody, past the test and control

lines, producing similar binding.

13

with different types of materials to choose

from, with varying treatments that can be

applied. Often, simply changing the type

of material or the treatment applied to it

can determine if the LFI runs successfully,

or if it fails. The materials can vary in size

and in overlap, with some tests being

long, others being thin, some with their

layers overlapping 2.5 mm, 1 mm. Each

test is different, with no simple,

streamlined pathway of development.

While developing an LFI, labs experiment

with these many variations, and expend

two valuable resources: time and funding.

Manufacturing developed tests requires

additional time and funding.

To streamline time and funding,

we propose reducing the layers of the LFI

from at least five layers to two: a plastic

backing card and treated paper.

Eliminating the sample pad, conjugate

pad, nitrocellulose membrane, absorbance

pad, and tape reduces the cost of materials

significantly. The time spent working

through each of these materials to find the

best fit can be dedicated to developing the

test in other areas. Manufacturing tests no

longer involves overlapping layers by

exact measurements and pressing the

overlaps together with tape or a cartridge,

but rather securing treated paper to a

plastic backing card. The expensive

machines used to stripe the tests are

replaced by a rollerball pen filled with

antibody ink. Treatment of each material

goes from a multitude of options to one:

modifying the paper to behave like

nitrocellulose membrane.

■ METHODS

Progression to “Sandwich Model”

Traditional LFI tests follow the

“sandwich model”, where a conjugated

antibody binds to the target biomarker.

This conjugated antibody-biomarker

complex binds to unconjugated antibody

that is secured to the nitrocellulose layer

as the test line. The test line constitutes

unconjugated antibodies that recognize

the conjugated antibody. The three

proteins (unconjugated antibody, antigen,

and conjugated antibody) form what is

called the “sandwich model” and is the

clinically relevant goal when developing a

diagnostic test.

To begin seLFI development, we

focused on the simplest version of this

clinically relevant model, termed “tier

one”. A tier one test was purposed to

answer two questions:

1. Would protein remain bound to the

modified paper test?

2. Could the bound protein withstand the

friction of fluid flow, or would

capillary action dislodge the bound

protein?

To answer these questions, we built a

minimalistic test with only one protein.

This was done by modifying printer

paper, either by nitration or by aldehyde-

functionalization, securing it to a plastic

backing card, and dividing it into test

strips. The test strips were striped with

conjugated antibody (⍺-hCG beta

antibody in goat, gold colloid, Fitzgerald)

and were subjected to capillary flow of

5% dry milk TBS-T buffer. Following the

tier one test, a tier two test was developed.

A tier two test built upon this

model, adding a protein to the test. Tier

two would answer the following question:

1. Could a test or control line capture

antibodies or other proteins as they

flow past in a buffer medium?

2. If so, will the captured proteins remain

bound as additional sample and buffer

clears the test or control lines?

Tier two tests were made with both

nitrated printer paper and aldehyde-

14

functionalized printer paper, then

compared to see which paper treatment

performed best. Modified printer paper

was striped with hCG protein (EMD

Millipore) and allowed to dry. 5% dry

milk TBS-T buffer spiked with conjugated

antibody (⍺-hCG beta antibody in goat,

gold colloid, Fitzgerald) was used to

induce capillary action. Since we stopped

using nitrated printer paper after tier two,

the details of this method are included in

the section dedicated to nitrated printer

paper.

Tier three was built as a final

phase preliminary test, answering the

standard questions:

1. Could a line of antibodies capture a

protein-conjugate complex?

2. Could a line of antibodies capture the

conjugated antibody?

At this point in our research, nitrated

printer paper had ceased to be a focus,

whereas aldehyde-functionalized printer

paper became our first choice for the

seLFI. Therefore, our tier three tests were

developed on aldehyde-functionalized

printer paper only, and not nitrated printer

paper. These tests were first made with

only a test line: a stripe of ⍺-hCG (⍺-hCG

alpha antibody in goat, MyBioSource)

that would be met with a complex of hCG

(EMD Millipore) and conjugated antibody

(⍺-hCG beta antibody in goat, gold

colloid, Fitzgerald). We then developed a

tier three test with only a control line: a

stripe of anti-host antibody (rabbit anti-

goat, Fitzgerald, MyBioSource) that

would be met with conjugated antibody

(⍺-hCG beta antibody in goat, gold

colloid, Fitzgerald). Since tier three was

our clinically relevant goal, the details are

outlined in the following section

dedicated to aldehyde-functionalized

printer paper.

Nitrated Printer Paper

Nitrocellulose membranes use

hydrophobic interactions to bind to

proteins. These hydrophobic interactions

are strong, and once bound, only extreme

measures can remove proteins from

nitrocellulose membrane. Therefore, when

nitrocellulose membrane is striped with

test and control antibodies, they remain

fixed. The surrounding active sites on the

membrane are blocked, and additional

proteins, such as those in a biological

sample, can flow through the membrane,

only binding to specific antibodies with

which they pair. Compared to

nitrocellulose membrane, paper is a low-

cost solution, but it must be modified to

secure proteins and keep them from

drifting as biological sample and buffer

flow through the paper. Nitration was the

first method explored to modify paper to

act like nitrocellulose membrane. Equal

parts of nitric and sulfuric acid were

carefully combined and stored at low

temperatures in a small glass vial. Printer

paper (75 g/m2, 92 bright, Target) was

attached to a plastic backing card and cut

into 5.0 mm x 34.5 mm strips. The chilled

acid mixture was pipetted across the

middle zone of the paper strips to create a

fine line (2.0-3.0 mm) that would hold the

test protein. After the acid mixture was

deposited and allowed to absorb, the

paper strips were run under continuous

water for five minutes, then incubated

with sodium bicarbonate and water under

shaking for an additional five minutes.

After neutralization, the nitrated paper

strips could dry at room temperature and

were stored in a sealed bag with desiccant

packs at room temperature.

In the first phase of research, hCG

was acting as the test protein, with

conjugated ⍺-hCG acting as the

biomarker spiking the buffer. This was a

“first-step” toward a more clinically

15

relevant “sandwich” model, with

antibodies as the test and control proteins,

and unconjugated proteins acting as the

biomarker. Initially, the tests were bare-

bones simple, involving only hCG and

conjugated ⍺-hCG. To prepare the test

strips, 1.0 μL of hCG at 1.0 mg/mL (the

test protein in this model) was pipetted

along the nitrated zone and allowed to dry

and bind. In order to prevent false

positives, the active sites of the nitrated

zone were blocked by pipetting 3 μL of a

5% solution of dry milk in TBS-T onto

and around the edges of the test zone.

After drying, the tests were ready to run.

Aldehyde-functionalized Printer Paper

Another successful method of

securing proteins to paper was developed

by Badu-Tawaiah et al. This method takes

cellulose and creates aldehyde groups

through a process of incubating the paper

in a solution of KIO4. Those aldehyde

groups can then form Schiff bases with

proteins that contact the paper, similar to

nitrocellulose. In this method, one sheet

of printer paper was placed in a 0.03 M

solution of KIO4. The paper was then

soaked for 2 hours at 65°C. After

treatment, the paper was rinsed three

times in fresh deionized water, and then

allowed to dry overnight at 35°C.

Once the paper was dry, antibodies

were added to the paper in test and control

lines. The test line antibodies and control

line antibodies were placed directly onto

the aldehyde functionalized paper using a

straight edge to guide the line and a 10 μL

syringe to apply the antibody (1.0

mg/mL). Later, syringes would be

replaced with antibody pens, which are

discussed hereafter. Since the entire

paper’s surface was functionalized after

treatment, the strips were blocked with

5% dry milk blocking buffer in water to

ensure that the paper would not have open

active sites. After the blocking, the paper

was dried, then placed onto a plastic

backing card. Once attached to the card,

the paper was cut into 5.0 mm x 34.5 mm

strips. The process of creating aldehyde

functionalized paper is simple and has the

potential to be delivered to the end user

ready to be striped and blocked.

To run the test strips, 90.0 µL of

5% dry milk solution in TBS-T was added

to 14 2.0 mL microcentrifuge tubes.

Seven of the tubes acted as positive

controls: six for the seLFI, and one for the

commercially available pregnancy test. To

these seven tubes, 1.0 µL of hCG (1.0

mg/mL) was added to each tube

separately and mixed gently with a

pipette. The other seven tubes acted as

negative controls, with six for the seLFI,

and the seventh for the commercially

available pregnancy test. 10.0 µL of gold

nanoparticle (AuNP) conjugated goat ⍺-

hCG antibodies were added to each tube

separately and mixed gently with a

pipette. After preparing these samples, test

strips were placed into each tube and

allowed to run via capillary action.

16

Antibody Pens

Aldehyde-functionalized paper

proved to be a strong, flexible,

inexpensive alternative to nitrocellulose.

Striping this paper with test and control

antibodies would traditionally be done

with an automated dispenser. However,

the goal of this project excluded the use of

expensive equipment and electricity, since

neither are currently feasible in low-

resource areas. Even other low-cost

options such as using a printer or a simple

X-Y plotter required electricity, and it

became clear that the target demographic

would require a more manual technique.

The best way to precisely deliver a

substance to a surface without electricity

is with a writing instrument. Research led

to filling felt-tipped markers, brush pens,

ballpoint pens, and paint pens with an

antibody solution “ink”, with the most

success of delivery being found in a

rollerball pen. These rollerball pens can

be filled with virtually any antibody ink

and are able to draw lines of antibodies

onto the aldehyde functionalized paper,

producing the control and test lines for

many different kinds of tests. We found

that J. Herbin refillable rollerball pens

(Figure 2) were best suited for delivering

antibody ink since they used a wick to

draw up the ink and used a refillable

piston cartridge to hold ink (Kaweco Mini

Piston Converter). This made it easy to

test, clean, and reuse the pen with

different types of antibodies. This method

of delivery promotes user-specific and

diverse diagnostic tests for a variety of

settings—including low-resource settings

and small team operations.

To test feasibility, the piston

cartridges were filled with rabbit ⍺-goat

antibodies and goat ⍺-hCG antibodies.

The control pen (goat ⍺-hCG) was used to

write on aldehyde functionalized paper

and nitrocellulose. Once the protein ink

had dried, the blots were blocked together

with 5% dry milk TBS-T buffer at 25℃

while shaking for 1 hour. After rinsing,

the blots were incubated in similar

conditions with TBS and 60 𝜇L hCG for 1

hour. After this incubation, the blots were

E D

Figure 2. Antibody Pen

Pictured is a sequential explosion of the

antibody rollerball pen. A) shows the pen

completely assembled and capped. B) shows the

pen uncapped, exposing the tip of the rollerball

pen. C) shows the body unscrewed from the tip

housing, showing how the piston cartridge

attaches securely to the tip housing, connecting to the wick. D) shows the depressed piston

cartridge detached from the tip housing,

exposing the faint outline of the white wick in

the tip housing. E) shows the same, but with the

piston cartridge completely extended. The piston

cartridge holds the antibody ink, which is

carried to the tip of the rollerball pen through a

wick when the two are properly connected. Each

part can be cleaned and reused with new or

different antibody when the user desires.

17

rinsed and incubated with 400 𝜇L AuNP-

conjugated goat ⍺-hCG antibodies under

similar conditions for 1 hour. The blots

were then rinsed and dried.

Additional tests were done to

determine the antibody pen’s longevity.

The antibody pen piston cartridge was

filled with 100 𝜇L of rabbit ⍺-goat

antibodies, secured to the pen, and then

was used to draw several straight lines of

antibody ink on a single sheet of paper

until the pen ran dry. The paper was then

blocked for 1 hour in 5% dry milk

solution in TBS-T, washed in deionized

water, then incubated with 400 𝜇L AuNP-

conjugated goat ⍺-hCG antibodies in TBS

under shaking for 1 hour. The paper was

removed, rinsed with deionized water, and

allowed to dry. The observed lines were

measured for length, determining the

approximate number of tests 100 𝜇L of

antibody in the antibody pen could make.

Testing in Urine

Although the seLFI had proven its

potential in a protein-spiked sample

buffer, a clinically relevant diagnostic

would need to perform in biological

samples. Testing the seLFI in biological

samples began with urine, due to

simplicity of training, collection, and

storage. To run the test strips, 90.0 µL of

urine were added to eight 2.0 mL

microcentrifuge tubes. 1.0 µL of hCG (1.0

mg/mL) was added to the first four tubes

separately and mixed gently with a

pipette, constituting our positive samples.

An additional positive control was

performed with a commercially available

pregnancy test, run with the fourth tube

according to the package directions. The

remaining four tubes were left without

hCG, constituting our negative samples.

An additional negative control was run

using a second commercially available

pregnancy test. To the remaining six tubes

(both positive and negative samples), 10.0

µL of AuNP-conjugated goat ⍺-hCG

antibodies were added to each tube

separately and mixed gently with a

pipette. After preparing these urine

samples, previously prepared aldehyde

functionalized test strips were placed into

each tube and allowed to run via capillary

action.

Testing in Serum

Another clinically relevant goal

was to test the seLFI in serum. 90.0 μL of

fetal bovine serum (FBS) were added to

eight 2.0 mL microcentrifuge tubes. 1.0

μL of hCG (1 mg/mL) was added to the

first four tubes separately and mixed

gently with a pipette, constituting the

positive samples. An additional positive

control was performed with a

commercially available pregnancy test,

run with the sample in the fourth tube

according to the package directions. The

remaining four tubes were left without

hCG, constituting the negative samples.

An additional negative control was run

using a second commercially available

pregnancy test. To the remaining six tubes

(both positive and negative samples), 10.0

μL of AuNP-conjugated goat α-hCG

antibodies were added to each tube

separately and mixed gently with a

pipette. After preparing these FBS

samples, previously prepared aldehyde

functionalized test strips were placed into

each tube and allowed to run via capillary

action.

Limit of Detection

A lateral flow immunoassay must

detect a biomarker; furthermore, it must

detect the correct levels of that biomarker

in the sample. Up until this point, we had

been loading a large amount of biomarker

18

(1 𝜇L of 1 mg/mL hCG) into sample

buffer to build a working preliminary

seLFI. To establish a working credible

seLFI, we had to determine the limit of

detection. The goal was to observe the

lowest level of biomarker the seLFI could

detect, especially in a way that the lowest

level could be read by the human eye. Our

desired outcome was for the seLFI to

detect the concentration of hCG in urine

during pregnancy, or 2.5 µg/mL.

To prepare the test runs, a serial

dilution was set up in six dilutions. 90 µL

of 5% dry milk solution in TBS-T was

added to seven 2 mL microcentrifuge

tubes. 10 µL of hCG (1 mg/mL

concentration, or 1:1) was added to and

mixed in the first dilution, resulting in

1:10 hCG concentration. 10 µL of

solution was removed from the first

dilution, added to the second

microcentrifuge tube and gently mixed,

resulting in a 1:102 hCG concentration.

This process was repeated until six tubes

contained hCG ranging from 1:10 to 1:106

concentrations. The final tube was left

without hCG and served as a negative

control.

This serial dilution was repeated

three times in additional microcentrifuge

tubes: two of the sets of serial dilutions

acted as replicates for the original

dilution, and one set of serial dilutions

would be used to run commercially

available pregnancy tests as a positive

control. These commercially available

tests were run according to the package

directions, using each tube separately for

a total of six positive control tests. The

final tube, having no hCG, served as the

negative control.

In the remaining three sets of

tubes, 10 µL of AuNP-conjugated goat α-

hCG was added to and gently mixed in

each separate dilution. After preparing

these dilution samples, previously

prepared aldehyde functionalized test

strips were placed into each tube and

allowed to run via capillary action.

This experimental set-up was

repeated to test the seLFI’s limit of

detection in urine, replacing the milk

chase buffer volume with the same

volume of urine, keeping all other

conditions the same.

■ RESULTS

Nitrated Printer Paper

In hopes of simplifying and

reducing the cost of the traditional LFI by

eliminating the sample pad, conjugate pad

and absorbent pad, a low-cost format that

was friendly to low-resource areas, the

single-sheet seLFI, was developed. This

experiment tested whether the seLFI

format could support stationary antibody-

antigen binding at the “test line”, like the

traditional LFI. After nitrating the printer

paper, striping with antibody, and

performing positive control seLFI tests,

some encouraging results followed. The

running buffer was entirely absorbed by

the seLFI test strip, and capillary action

took the conjugated ⍺-hCG past the test

line of hCG protein, which was bound to

the modified area of the test strip. Binding

between antigen and conjugated antibody

occurred, similar to a traditional LFI

sandwich format. Figure 3 shows the

genuine promise of performing a

diagnostic test using only one layer of

treated printer paper. The nitrated zone

acted as a pseudo test line (the first

antigen and conjugated antibody model

19

used was not a true test line), presenting a

visual positive result. These visual results

indicated that the modified areas of the

paper functioned akin to nitrocellulose;

stationary test lines on printer paper can

be maintained even after the introduction

of running buffer and sample flow.

Despite the preliminary success of

these elementary test results, several

points were concerning. Visually, the

results were diffuse and nonuniform,

making it difficult for technicians to

distinguish a positive result from a

negative result. Each test produced a

dramatically different result, presenting a

challenge to the tests’ reliability. On

several of the tests, the nitrated zone

behaved as a hydrophobic barrier,

encouraging the sample to flow around

the zone, rather than directly through it.

Notwithstanding the practice of blocking

around the nitrated area after striping with

antibody, negative tests were inconsistent

(not pictured). Outside of test

performance, test manufacture was risky

and inconvenient, due to the use of strong

acids. With these challenges and dangers,

it was decided to pursue an alternative to

nitrated printer paper.

Aldehyde-functionalized Printer Paper

In a safe and convenient

alternative to nitrated printer paper, the

method of Badu-Tawaiah et al. was used

to aldehyde-functionalize printer paper.

Printer paper was treated with KIO4,

separate lines of test and control

antibodies were striped, and the paper was

blocked and cut into test strips. Running

the tests with positive and negative

samples (5% dry milk solution in TBS-T

with and without hCG) produced clear,

uniform, and legible results (Figure 4).

These results demonstrate the seLFI’s

ability to compare to traditional LFI

results, in speed, reliability, and

specificity.

In the positive seLFI results

(Figure 4: A), the milk chase buffer,

AuNP ⍺-hCG, and hCG were drawn up

via capillary action, crossing the entire

test, including the test and control lines.

The hCG/AuNP ⍺-hCG complex

remained bound to the test line, and the

AuNP ⍺-hCG remained bound to the

control line, resulting in the visual result

of two pink lines. In the negative seLFI

results (Figure 4: B), the chase buffer and

AuNP ⍺-hCG ran through the entire test,

crossing both the test and control lines.

There was no hCG/AuNP ⍺-hCG

complex, so nothing bound to the test line,

but the AuNP ⍺-hCG remained bound to

the control line, indicating a valid test.

Figure 3. seLFI with Nitration Method

Nitrated seLFI test strips were made with

printer paper and plastic backing cards. Using

a 1:1 solution of nitric and sulfuric acids, the

printer paper was chemically modified to bind

proteins securely. After being cut into 5.0 mm

x 34.5 mm strips, 1 μL of 1.0 mg/mL hCG was

deposited on and bound to the nitrated middle

portion of the strips, and the remaining area of the strips were blocked with a 5% dry milk

solution in TBS-T to prevent non-specific

binding. AuNP-conjugated ⍺-hCG was run as

the “sample”. After the tests had run, red lines

could be seen, indicating binding between

hCG and ⍺-hCG. The weaknesses of this

model include impracticality of assembly,

irregular binding, and a slight hydrophobic

barrier that appears at the nitrated zone,

preventing complete flow.

20

This yielded the visual result of one pink

line. The commercially available

pregnancy tests (Figure 4: C) show the

negative control (left) and positive control

(right). These tests show an intensity of

color in the results and a low background

that surpassed the seLFI tests. Though the

seLFI tests exhibit potential in accuracy,

specificity, and visual readability,

optimization was needed.

Antibody Pen

The antibody pen was used to

draw letters on to aldehyde-functionalized

paper using goat ⍺-hCG antibody ink.

These markings could be seen as wet lines

until they dried; after drying, the writing

disappeared entirely. The consequent

steps in the experiment—blocking,

incubating with hCG, then incubating

with AuNP ⍺-hCG—were intended to

facilitate a “sandwich” complex between

the invisible goat ⍺-hCG that was drawn

onto the paper, the suspended hCG, and

the suspended AuNP ⍺-hCG. This test

would determine if the antibody pen

allowed for proteins to pass through the

system without being damaged, and if the

delivery would be detailed, even, and

precise (Figure 5). Figure 4. Aldehyde-functionalized seLFI Strips. Half-strips were prepared with a 0.03 M

solution of KIO4 and heat, then striped with

lines of control (top line) and test (bottom line)

antibodies. The striped paper was blocked and

cut into strips, then the individual tests were

run. A) Positive tests were performed with 30

μL of 5% dry milk buffer, 10 μL of gold

nanoparticle conjugated ⍺-hCG, and 1 μL of 1

mg/mL hCG. Results are clear, with both the control and test lines showing specific binding.

B) In turn, the negative tests were performed

under similar conditions, with the exception of

hCG. As expected, only the control line

showed specific binding, with a blank test line

indicating the absence of hCG. C) Additional

commercially produced positive controls were

run under similar conditions to verify the

positive and negative seLFIs.

Figure 5. Antibody Pen Precision and Efficacy To determine if the antibody pens would write

evenly and precisely without damaging the

antibodies in the ink, the piston cartridge was

filled with rabbit ⍺-goat antibodies. It was then

used to write on aldehyde functionalized paper

and nitrocellulose. Once the protein ink had dried, the blots were blocked together with 5%

dry milk TBS-T buffer at 25℃ while shaking

for 1 hour. After rinsing, the blots were

incubated in similar conditions with TBS and 60

𝜇L hCG for 1 hour. After this incubation, the

blots were rinsed and incubated with 400 𝜇L

AuNP-conjugated goat ⍺-hCG antibodies under

similar conditions for 1 hour. The blots were

then rinsed and dried. A and B show results on nitrocellulose, while C and D show results on

aldehyde-functionalized paper. A and C were

written by technician A, and B and D were

written by technician B.

21

These completed tests presented

clear, uniform, legible lines. For both

aldehyde-functionalized paper and

nitrocellulose paper, the antibody delivery

is so detailed that the differences in

handwriting between the two technicians

can be identified. When run under

identical conditions, the aldehyde

functionalized paper produced a clearer,

darker result than the nitrocellulose paper.

The pen has been licensed to DCN

Diagnostics and is currently being

developed to be included in their LFI

starter kits.

The second test determined the

lifetime of 100 μL of antibody ink in the

pen (Figure 6). After drawing, blocking,

incubating, and drying, the lines were

measured to project the total number of

tests 100 μL of antibody could produce. A

total of 5,588 mm was measured before

the integrity of the line began to break up,

leaving the rest unmeasured for the sake

of quality control. This length would

result in approximately 1,117 tests. These

tests indicate that the antibody pen is

feasible with the low-cost, machine-free

manufacturing system required by a low-

resource lab or clinic.

Testing in Urine

Testing the seLFI in urine would

expand the platform from a research

relevant diagnostic to a clinically relevant

diagnostic. After running the hCG seLFI

tests in milk buffer successfully, we

reproduced the testing procedures for

testing in urine, simply replacing the milk

chase buffer with the same amount of

urine (Figure 7). Production of the tests

and the amounts of hCG and AuNP ⍺-

hCG remained the same as the tests run in

milk chase buffer. The tests performed

Figure 7. Aldehyde-functionalized seLFI Strips in Urine Half-strips were prepared according to the

aldehyde-functionalization method. Positive

tests (left) were performed with 90 μL of urine,

10 μL of AuNP ⍺-hCG, and 1 μL of 1 mg/mL hCG. Results are clear, with both the control

and test lines showing specific binding.

Negative tests (right) were performed under

similar conditions, without the addition of

hCG. As expected, only the control line

showed specific binding, with a blank test line

indicating the absence of hCG. Commercially

available (center) positive and negative

controls were run under according to package

directions to validate the seLFI results.

Figure 6. Antibody Pen Extinction Test This test was designed to better understand the

lifetime of antibody ink in the pen. 100 𝜇L of

rabbit ⍺-goat antibodies were used to fill the

piston cartridge, then drawn out in straight

lines until the antibody ink ran out. The paper

was then blocked for 1 hour in 5% dry milk

solution in TBS-T, washed in deionized water,

then incubated with 400 𝜇L AuNP-conjugated

goat ⍺-hCG antibodies in TBS under shaking

for 1 hour. The paper was removed, rinsed

with deionized water, and allowed to dry. The

lines were later measured and broken down

into an approximate yield of tests.

22

exceptionally well, revealing clear test

and control lines in the positive samples,

and clear control lines in the negative

samples. The tests ran significantly faster

in urine than they did in milk chase

buffer: a quick 2 minutes compared to 20

minutes.

Testing in Serum

With encouraging results from

testing in urine samples, the next step was

to test the seLFI in fetal bovine serum

(Figure 8). Like the urine tests, we simply

replaced the milk chase buffer with the

same amount of FBS, keeping all other

variables consistent. These tests still

showed promise, presenting consistent

positive and negative results, and taking

only about 2 minutes to complete.

Figure 8. Aldehyde-functionalized seLFI Strips in Fetal Bovine Serum Half-strips were prepared according to the

aldehyde-functionalization method. Positive

tests (left) were performed with 90 μL of fetal

bovine serum, 10 μL of AuNP ⍺-hCG, and 1

μL of 1 mg/mL hCG. Results are clear, with

both the control and test lines showing specific

binding. Negative tests (right) were performed

under similar conditions, without the addition

of hCG. As expected, only the control line

showed specific binding, with a blank test line indicating the absence of hCG. Commercially

available (center) positive and negative

controls were run under according to package

directions to validate the seLFI results.

23

Limit of Detection

Finding the seLFI’s limit of

detection would allow for a side-by-side

comparison of the seLFI’s sensitivity and

the sensitivity of commercially available

tests. Ideally, the seLFI would be able to

detect hCG at the level of 1:104 in order to

detect 2.5 ng/mL of hCG in a positive

urine sample. The seLFI consistently

bound hCG at concentrations of 1:102 to

1:103, with unreliable, faint binding at

1:104 (Figure 9). Interestingly, a

Figure 9. Aldehyde-functionalized seLFI Limit of Detection Half-strips were prepared according to the aldehyde-functionalization method. All samples were

prepared in similar fashion, with 90 L of sample in each tube and 10 L of hCG used to make a serial

dilution of values from 1:10 to 1:106. These were repeated three times to make a total of four runs: three

dedicated to testing the seLFI, and one dedicated as a control with commercially available pregnancy

tests. A-C show results with milk buffer, with D as the control. E-G show results with urine, with H as

the control.

A

B

C

D

E

F

G

H

24

concentration of hCG at 1:10 resulted in

weak test lines as well. Acknowledging

the inconsistent binding at 1:10 and 1:104

and beyond, we determine the seLFI limit

of detection to be between 1:102 and

1:103. These results were consistent across

all three trials for both milk chase buffer

(Figure 9 A-D) and urine samples

(Figure 9 E-H).

■ DISCUSSION

The aldehyde-functionalized paper

is a low-cost, safe, flexible replacement

for nitrocellulose.

For small labs or clinics with low funding

or few resources, purchasing

manufacturing machinery to make tests

for labs, or purchasing pre-made tests for

clinics is financially challenging. Having

an electricity-free instrument that could be

used to make diagnostic tests for research

or for clinical use would serve those

populations well. The antibody pen

delivers similar results at a small scale.

The low price and simple maintenance

and use of the pen lends itself well to

diagnostic development, as well as fast

test manufacture for in-field use.

Diagnostic tests use biological

samples to detect the presence of

biomarkers in the sample, becoming a tool

to aid in diagnosis and treatment of

diseases and conditions. Biological

samples range from whole blood to

serum, tears to saliva, urine to feces, and

many more. The ability to detect

biomarkers in biological samples was key

to the success of the seLFI, allowing it to

function as a diagnostic tool. Due to the

presence of hCG in urine during

pregnancy, it was encouraging to observe

the seLFI’s results when tested in urine. A

short running time with clear, legible

results opens the possibility of developing

a test for a different biomarker in a

different sample type. Although urine has

a connection to the biomarker used in this

study, we chose to test for hCG added into

serum during sample preparation. The

purpose of these tests was to explore the

possibility of testing in different types of

samples, while still working with a

biomarker and antibodies that we were

familiar with. The results of the seLFI

serum tests were likewise encouraging

and led us to plan to test the seLFI in

whole blood and saliva. The goal is to

prove the seLFI platform in as many

biological samples as possible, in order to

establish its potential as a new,

innovative, and versatile diagnostic

platform.

In order to be an effective test, the

seLFI’s sensitivity would need to achieve

a limit of detection of 1:104. In order to

improve sensitivity, a few avenues will be

pursued. The first would be to simply

increase the amount of AuNP ⍺-hCG used

while running the test. This would

saturate the available hCG, increasing the

intensity of the test line, despite the lower

levels of protein in the sample. Another

avenue involves antibody-antigen

pairings. With a fade-out at 1:104

concentration of hCG, the issue is

between the ⍺-hCG, hCG, and AuNP ⍺-

hCG sandwich complex. Finding a variety

of these proteins and would allow us to

set up a protein matrix. This would

involve a series of blot tests, each with the

same ⍺-hCG variety blotted onto a small

square of aldehyde-functionalized paper,

then blocked separately. Each of the blots

would then be incubated with the same

variety of hCG, then washed, then each

blot would be incubated with a different

variety of AuNP ⍺-hCG. These blot

matrices would be repeated, adjusting the

varieties of ⍺-hCG, hCG, and AuNP ⍺-

hCG appropriately until an ideal match is

25

discovered. This ideal match would

indicate the proteins that have the highest

binding affinity, increasing the amount of

hCG binding to both ⍺-hCG and AuNP ⍺-

hCG, strengthening the intensity of the

test line. An alternative avenue would be

to increase the concentration of ⍺-hCG on

the test line, so as to capture more hCG-

AuNP ⍺-hCG complexes, raising the level

of sensitivity.

The possibility of a false-negative

would need to be addressed, considering

the weak, unreliable binding at

concentrations of hCG at 1:10. To prevent

a false-negative, protein interactions at

concentrations of 1:10 would need to be

studied and understood more thoroughly

in order to plan experiments and adjust

the test and its procedures.

26

REFERENCES

A. Badu-Tawiah, S. Lathwal, K. Kaastrup, M. Al-Sayah, D. C. Christodouleas, B.

S. Smith, G. M. Whitesides, and H. Sikes. 2015. “Polymerization-based Signal

Amplification for Paper-Based Immunoassays.” Lab on a Chip, 15, Pp. 655-659.

Please see supplementary information in A. Badu-Tawiah’s publication:

http://www.rsc.org/suppdata/lc/c4/c4lc01239a/c4lc01239a1.pdf

T. Liang, R. Robinson, J. Houghtaling, G. Fridley, S. A. Ramsey, E. Fu. 2016.

“Investigation of Reagent Delivery Formats in a Multivalent Malaria Sandwich

Immunoassay and Implications for Assay Performance.” Analytical Chemistry, 88, Pp.

2311-2320